AI’s popularity and influence continue to surge in corporate America, as evident in the remarkable increase of references to AI and related terms during investor calls. Compared to a year ago, these mentions have more than doubled, highlighting the sustained interest in and focus on artificial intelligence within the business world. Alphabet, Amazon, Meta, and Microsoft alone mentioned it for a total of 168 times as some mention AI to bolster their surprise earnings and revenue to take attention away from their lukewarm guidance.
This is especially ironic since during the same time last year the hot topic was an increasing trend in unionization with Amazon’s first union in the US forming at their New York warehouses in April of last year. While this year we are seeing a lot of chatter that AI chatbots will be replacing most of our favorite Wendy’s drive thru attendants in the near future (pilot tests start in June according to recent news). However, before the pitchforks are raised against AI let us first understand if the chatbot can really “terk err jerbs”! In a nutshell, the way AI works is the development of computer systems capable of performing tasks that typically require human intelligence such as jotting down how many Frosty one would want for this upcoming summer (as long as Boston Dynamics hasn’t teamed up with ChatGPT yet to build a Janitor robot the busboys are safe). To date AI’s advancements in recent years have mostly come in the form of Machine Learning and Deep Learning, Natural Language Processing, Computer Vision, Reinforcement Learning, and Generative AI to name a few. The last one being quite controversial as of late since this is the subfield of AI that generates content and has been used in a variety of ways such as creating art to writing homework and even creating an endless stream of Seinfeld based jokes.
AI has undoubtedly made significant progress, but it still faces some key challenges. Data dependency is a key challenge, as AI models require large, diverse, and unbiased datasets for effective training. Another challenge, especially for the use of AI (especially in Finance) is the lack of interpretability and auditability, as complex AI models often lack transparency in their decision-making processes. Adversarial attacks also pose a threat to the robustness and security of AI systems as it is still vulnerable to data manipulation that can alter the outputs generated. Generalization of AI’s knowledge and applying it to new scenarios is still something that AI is having trouble hurdling. The final limitation that we’ll tackle is resource intensiveness of AI, both in terms of computational requirements and energy consumption, as the computing power to implement a Generative Pre-Trained Transformer (ChatGPT) is significantly more costly than a chatbot taking drive-thru orders.
While the challenges for AI to fully replace human jobs are currently insurmountable, it is crucial to acknowledge that AI will continue its progression towards achieving human-level intelligence. Instead of fearing its potential impact, we should embrace AI as a tool for our own advancement and improvement. Recognizing AI’s role as a complementary tool rather than a threat allows us to harness its capabilities and collaborate with this technology to drive positive outcomes, enhance our own abilities, and keep our jobs.